University of Nebraska - Lincoln DigitalCommons@University of Nebraska - Lincoln

Faculty Publications: Department of Entomology Entomology, Department of

7-2013 Waveform Library for Chinch Bugs (: : Blissidae): Characterization of Electrical Penetration Graph Waveforms at Multiple Input Impedances Elaine A. Backus USDA-ARS, [email protected]

Murugesan Rangasamy University of Florida

Mitchell Stamm University of Nebraska-Lincoln

Heather J. McAuslane University of Florida, [email protected]

Ron Cherry University of Florida, [email protected]

Follow this and additional works at: http://digitalcommons.unl.edu/entomologyfacpub Part of the Entomology Commons, and the Research Methods in Life Sciences Commons

Backus, Elaine A.; Rangasamy, Murugesan; Stamm, Mitchell; McAuslane, Heather J.; and Cherry, Ron, "Waveform Library for Chinch Bugs (Hemiptera: Heteroptera: Blissidae): Characterization of Electrical Penetration Graph Waveforms at Multiple Input Impedances" (2013). Faculty Publications: Department of Entomology. 617. http://digitalcommons.unl.edu/entomologyfacpub/617

This Article is brought to you for free and open access by the Entomology, Department of at DigitalCommons@University of Nebraska - Lincoln. It has been accepted for inclusion in Faculty Publications: Department of Entomology by an authorized administrator of DigitalCommons@University of Nebraska - Lincoln. BEHAVIOR Waveform Library for Chinch Bugs (Hemiptera: Heteroptera: Blissidae): Characterization of Electrical Penetration Graph Waveforms at Multiple Input Impedances

ELAINE A. BACKUS,1,2 MURUGESAN RANGASAMY,3,4 MITCHELL STAMM,5 3 6 HEATHER J. MCAUSLANE, AND RON CHERRY

Ann. Entomol. Soc. Am. 106(4): 524Ð539 (2013); DOI: http://dx.doi.org/10.1603/AN13015 ABSTRACT Electrical penetration graph (EPG) monitoring has been used extensively to elucidate mechanisms of resistance in plants to herbivores with piercing-sucking mouthparts. Charac- terization of waveforms produced by during stylet probing is essential to the application of this technology. In the studies described herein, a four-channel Backus and Bennett ACÐDC monitor was used to characterize EPG waveforms produced by adults of two economically important chinch bug species: southern chinch bug, insularis Barber, feeding on St. Augustinegrass, and western chinch bug, Barber, feeding on buffalograss. This is only the third time a heteropteran species has been recorded by using EPG; it is also the Þrst recording of adult heteropterans, and the Þrst of Blissidae. Probing of chinch bugs was recorded with either AC or DC applied voltage, no applied voltage, or voltage switched between AC and DC mid-recording, at input impedances ranging from 106 to 1010 ⍀, plus 1013 ⍀, to develop a waveform library. Waveforms exhibited by western and southern chinch bugs were similar, and both showed long periods of putative pathway and ingestion phases (typical of salivary sheath feeders) interspersed with shorter phases, termed transitional J wave and interruption. The J wave is suspected to be an X wave, that is, in EPG parlance, a stereotypical transition waveform that marks contact with a preferred ingestion tissue. The ßexibility of using multiple input impedances with the ACÐDC monitor was valuable for determining the electrical origin (resistance vs. electromotive force components) of the chinch bug waveforms. It was concluded that an input impedance of 107 ⍀, with either DC or AC applied voltage, is optimal to detect all resistance- and electromotive force-component waveforms produced during chinch bug probing. Knowledge of electrical origins suggested hypothesized biological meanings of the waveforms, before time-intensive future correlation experiments by using histology, microscopy, and other techniques.

KEY WORDS Insecta, Blissus, EPG, feeding, stylet penetration

Chinch bugs within the genus Blissus have long been pest of St. Augustinegrass, Stenotaphrum secundatum recognized as important economic pests of cultivated (Walter) Kuntze, a well-adapted and widely planted grasses within the United States. Two species, in par- turfgrass species in the southern United States (Kerr ticular, have gained notoriety as pests of turf and 1966, Reinert and Kerr 1973). The southern chinch pasture grasses. The southern chinch bug, Blissus in- bugÕs North American geographic distribution extends sularis Barber (Hemiptera: Blissidae), is the major from South Carolina to Florida, through the Gulf Coast states into Texas and Oklahoma, and also into The use of trade, Þrm, or corporation names in this publication is California and Hawaii (Vittum et al. 1999). Its host for information and convenience of the reader. Such use does not range encompasses important turfgrasses such as Ber- constitute an ofÞcial endorsement or approval by the United States muda grass, Cynodon spp., and centipedegrass, Er- Department of Agriculture or the Agricultural Research Service of emochloa ophiuroides (Slater 1976), but St. August- any product or service to the exclusion of others that may be suitable. USDA is an equal opportunity provider and employer. inegrass is highly preferred (Reinert et al. 2011). A 1 Crop Diseases, Pests and Genetics Research Unit, USDA Agri- closely related species, the western chinch bug, Blissus cultural Research Service, San Joaquin Valley Agricultural Sciences occiduus Barber, is primarily an economic pest of buf- Center, 9611 South Riverbend Ave., Parlier, CA 93648. falograss, Buchloe dactyloides (Nuttall) Englemann, a 2 Corresponding author, e-mail: [email protected]. ¨ 3 Department of Entomology and Nematology, University of Flor- low maintenance and relatively pest-free new turf- ida, Gainesville, FL 32611-0620. grass for the midwestern United States (Baxendale et 4 Present address: Dow AgroSciences, LLC, Indianapolis, IN 46268. al. 1999). The western chinch bug has a broad host 5 Department of Entomology, University of NebraskaÐLincoln, Lin- range that includes turfgrasses; agronomic crops such coln, NE 68583. 6 Everglades Research and Education Center, University of Florida, as wheat, barley, rye, and sorghum; and native grasses Belle Glade, FL 33430. (Eickhoff et al. 2004). It is found predominantly in the

0013-8746/13/0524Ð0539$04.00/0 ᭧ 2013 Entomological Society of America July 2013 BACKUS ET AL.: WAVEFORM LIBRARY FOR CHINCH BUGS 525 western United States and Canada (Baxendale et al. forms can be characterized based on frequency, am- 1999). plitude, and appearance, and their electrical origin can Direct feeding damage by these insects consists of be attributed to changes in resistance (i.e., R compo- discoloration of the affected tissue (Baxendale et al. nent of the waveform) or production of biopotentials 1999), with increasing areas of yellow or brown turf, (i.e., electromotive forces, or emf component), or a and eventual death of large patches or even entire mixture of both. The biological meanings of many R lawns or pastures (Reinert and Kerr 1973). Host plant versus emf electrical origins are now clearly under- resistance to these grass-feeding chinch bugs has been stood (Walker 2000). For example, emf components in investigated as a way to manage these pests in a more waveforms are caused by plant cell membrane break- environmentally friendly manner with reduced insec- ages by stylets (termed intracellular punctures), or ticide application. Categories of resistance to their streaming potentials formed by ßuid movements respective chinch bug pests have been elucidated in through capillary tubes like the food canal. R compo- several cultivars of St. Augustinegrass and buffalograss nents are derived from electrical conductivity of sa- (Reinert et al. 1986; Crocker et al. 1989; HengÐMoss et liva, electrical resistance to current ßow from move- al. 2002, 2003; Rangasamy et al. 2006; Eickhoff et al. ment of valves (e.g., precibarial valve and cibarial 2008), but the underlying mechanisms of resistance diaphragm) in the hemipteran head, or depth of the are still unknown. The key to discovering these mech- stylets relatively to the interior of the plant, which is anisms will be a better understanding of chinch bug hypothesized to conduct higher voltages of applied feeding. signal (Walker 2000, Backus and Bennett 2009). The three Blissus species whose feeding effects on Certain hemipteran taxa, mostly , whiteßies, plants have been examined histologically are catego- and other sternorrhynchans, have been monitored rized as salivary sheath feeders, as sheaths of gelling primarily by using direct current (DC) monitors, saliva were found in plants fed on by chinch bugs. The where DC is applied to the feeding substrate (Tjallin- sheath hardened around the chinch bug stylets as they gii 1978) at a Þxed ampliÞer sensitivity (or input re- advanced within the plant to the target tissue. Painter sistance/impedance, Ri) of 109 ⍀; output from these (1928), probably studying the common chinch bug, monitors emphasizes emf components of waveforms. leucopterus (Say), noted that a Other hemipterans, in particular larger insects and chinch bug feeding on corn and sorghum leaves laid those thought to be sensitive to DC, such as heterop- down a salivary sheath as its stylets passed intracellu- terans and auchenorrhynchans (sharpshooter leaf- larly on the way to the target phloem tissue. He at- hoppers), have been recorded mainly with alternating tributed observed plant damage to the removal of current (AC) monitors (Backus 1994, Backus and photosynthate from phloem, and possibly water from Bennett 2009), formerly with a Þxed Ri of 106 ⍀; the xylem, as well as clogging of the sieve tubes and output from these monitors emphasized primarily R tracheids with salivary sheath material (Painter 1928). components. Recordings of the same species of insect Salivary sheaths of western and southern chinch bug with both AC and DC monitors often produced very are frequently branched as they move through the leaf different waveforms, confusing researchers for many tissue (Anderson et al. 2006, Rangasamy et al. 2009), years (Backus 1994). A new EPG monitor design was and in the case of the western chinch bug, terminate recently developed that allows recording of insects primarily in the vascular tissue (Anderson et al. 2006). with either AC or DC applied signal as well as select- The target tissue of southern chinch bug stylets has not able Ri levels (Backus and Bennett 2009), and repre- been deÞnitively identiÞed; however, stylets routinely sents the latest reÞnement in the quest for a universal pass through up to Þve leaf sheaths as the insects EPG monitor (Backus et al. 2000). attempt to reach the youngest meristematic tissue The new ACÐDC four-channel monitor is similar in within a leaf shoot (Rangasamy et al. 2009). circuitry to the ACÐDC Correlation Monitor (Backus The use of electronic feeding monitors (McLean and Bennett 2009). Signal can be applied to the plant and Kinsey 1964) is the most rigorous method to in- as AC, DC, or not at all (0 V). Ri is fully selectable from vestigate the feeding behavior of piercing-sucking ar- 106 to 1010 ⍀, plus 1013 ⍀, permitting a researcher to thropods (Walker 2000, van Helden and Tjallingii match the ampliÞer sensitivity to the inherent resis- 2000). Stylet penetration behavior (or probing) of tance (Ra) of the insect, thus enabling detection of the piercing-sucking insects in the order Hemiptera has widest array of waveforms for each species. By switch- been studied for more than four decades by using ing among Ri levels, one can identify waveforms com- electronic (renamed electrical penetration graph posed entirely of R components (those seen at 106 ⍀ [EPG]; Tjallingii 1985) monitoring, in which the Ri) to entirely emf (those seen at 1013 ⍀ Ri). For mouthparts of a wire-tethered insect complete the waveforms composed of a mixture of R and emf, electrical circuit between a plant (substrate) and an changes in waveform amplitude as Ri is increased or EPG monitor. Waveforms produced by changes in decreased reveal whether the primary component is R resistance within the circuit plus generation of volt- or emf, following the RÐemf responsiveness curve. ages within the insect or plant (biopotentials) have Thus, an understanding of a waveformÕs R and/or emf been correlated with movement of the insectÕs stylets components can provide electrical evidence of the and other activities, such as salivation and ingestion biological processes occurring inside the plant and (Walker 2000), rendering visible the otherwise invis- insect during recording of these waveforms (Walker ible act of probing by stylet-feeding insects. Wave- 2000, Backus and Bennett 2009). 526 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 4

To use EPG to investigate the mechanisms of re- Western chinch bugs were collected from buffa- sistance to chinch bugs in varieties of St. Augustine- lograss stands at the University of NebraskaÐLincoln grass and buffalograss, one must Þrst characterize the (Lancaster County, NE) by vacuuming the soil sur- waveforms produced. However, no EPG recording has face with an ECHO Shred ÔNЈ Vac (model No. ES-250; been done on any Blissus species, or any species within ECHO Inc., Lake Zurich, IL). Chinch bugs were sifted the blissid family. In fact, EPG recordings of only two from samples through a 2-mm mesh screen (HengÐ heteropteran species have been previously published. Moss et al. 2002) and brachypterous adults were col- Feeding of Þrst- and second-instar-nymphs of the lected with an aspirator. squash bug, Anasa tristis (De Geer) (Hemiptera: Co- Plugs of ÔPrestigeÕ buffalograss (10.6 cm in diameter reidae), a salivary sheath feeder, was recorded with an by 8 cm in depth) were removed from Þeld plots at the older-generation DC monitor (Bonjour et al. 1991, John Seaton Anderson Turf and Ornamental Research Cook and Neal 1999) with Þxed Ri of 109 ⍀. Wave- Facility, University of Nebraska Agricultural Research forms from third instars of a cell rupture-feeding het- and Development Center. Buffalograss stolons were eropteran, Lygus hesperus (Hemiptera: ) vegetatively propagated in the greenhouse in cone- (Cline and Backus 2002, Backus et al. 2007), were tainers (3.8 cm in diameter by 21 cm in depth) (Stu- characterized by using an AC “Missouri monitor” ewe & Sons Inc., Corvallis, OR) containing a potting (type 2.2, Backus and Bennett 1992) with Þxed Ri of mixture of sandÐsoilÐpeatÐperlite (0.66:0.33:1:1) 106 ⍀. (Eickhoff et al. 2006). The plants were maintained In the current study, we characterized the EPG in a greenhouse (25ЊC, 75% RH, and a photoperiod waveforms of adults of two chinch bug species, in a of 16:8 L:D h) under 400-W high-intensity discharge third family of heteropterans to be studied by EPG lamps. Fully established sprigs (4Ð5 wk after plant- (Blissidae), by using the new ACÐDC four-channel ing) were then used for EPG recordings. monitor (Backus and Bennett 2009). Waveform fam- Wiring and Acclimating of the Insects. Female ilies and types were described for insects feeding on southern chinch bugs were removed from the colony plants supplied with AC, DC, or 0 V applied signal, at and starved for 3 h before wiring. A 12.7-␮m (in di- sequential Ri levels of 106 to 1013 ⍀, providing indi- ameter; sold as 0.0005 in) gold wire 1Ð2 cm in length cations of the relative amount of R and emf, and thus (Sigmund Cohn Corp., Mt. Vernon, NY) was glued to allowing us to hypothesize the stylet activities and the pronotum of the insect by using silver conductive biological processes represented by the characterized paint (Ladd Industries, Burlington, VT). The insect waveforms. Herein, we introduce a new term for EPG was held immobile during application of the wire science: waveform library, which is deÞned as a col- tether by placing its abdomen in a Pasteur pipette tip lection of signal output traces, measurements, and under low suction. After wiring, the insects were al- calculations from EPG recordings of the same species lowed to rest for a 1-h period on a moist paper towel at multiple input impedances. to acclimate to the wire. Western chinch bugs (sex undetermined) were starved for 17 h overnight, similarly immobilized, and then a gold wire (25 ␮m in diameter [sold as 0.001 in] Materials and Methods by 1Ð2 cm in length) was adhered to the pronotum of Plants and Insects. Southern chinch bugs were col- the chinch bug with a small drop of silver conductive lected from St. Augustinegrass lawns in Royal Palm glue, composed of white glue (ElmerÕs Products Inc., Beach (Palm Beach County, FL) by using a modiÞed Westerville, OH), water, and silver ßake (Inframat WeedEater Barracuda blower/vacuum (Poulan/ Corp., Manchester, CT) in a 1:1:1 (vol:vol:wt) ratio. Weedeater, Shreveport, LA) (Cherry 2001). Fifth in- EPG Equipment and Data Acquisition. A four- stars were separated from the debris, placed in 20-liter channel ACÐDC EPG monitor (Backus and Bennett plastic buckets with a moist paper towel covering the 2009; custom-built by the late W. H. Bennett) was used bottom, and provided with fresh ÔFloratamÕ St. Au- for recording the feeding behavior of four chinch bugs gustinegrass stolons. Insects were kept in a climate- simultaneously on their respective host plants. Plants, controlled room (26ЊC, 40% relative humidity [RH], a insects, and the head-stage ampliÞers were all placed photoperiod of 14:10 L:D h) at the University of Flor- in a Faraday cage, constructed of 1.27-cm mesh gal- ida, Gainesville. Three- to 5-d-old adult females were vanized hardware metal cloth (LG Sourcing Inc., N. used in the experiments. Wilkesboro, NC). Pots containing test plants were Rooted single node cuttings of Floratam were placed on wooden blocks to electrically isolate them grown in plastic pots (12.5 cm in diameter by 12.5 cm from the Faraday cage. A selected stolon from each in depth) in a potting medium composed of equal parts grass plant was laid down along a Plexiglas stage mea- Þne sand (Florida Rock Inc., Jacksonville, FL) and suring 20 cm in length [by] 5 cm in width and Þxed in Fafard Growing Media (Mix No. 2) (Conrad Fafard, place by using strips of ParaÞlm (Pechiney Plastic Agawam, MA). Plants were fertilized with Osmocote Packaging, Menasha, WI). The Plexiglas stage was Plus 15-9-12 (ScottsÐSierra Horticultural Products held horizontally by an alligator clip connected to a Company, Marysville, OH) at the rate of 5 g/kg of “helping hand” holder (van Sickle Electronics, St. growing medium. Plants used as feeding substrate in Louis, MO). The use of the horizontal stage allowed the experiments were 6Ð8 wk old and had at least one insects that had left the plant the opportunity to return stolon 30Ð40 cm in length. by walking to the plant (unless they fell off the stage July 2013 BACKUS ET AL.: WAVEFORM LIBRARY FOR CHINCH BUGS 527 and hung suspended from their wire tether). A 24- was achieved, to provide waveforms for comparison gauge copper wire (3 cm in length) inserted into the with those of southern chinch bugs. The experiment soil served as the plant electrode. The changes in was terminated as soon as any chinch bug was no electrical resistance as well as biopotentials during longer performing J-I2. stylet probing were ampliÞed, rectiÞed, and digitized In the second experiment, three insects at a time at 100 Hz by using a DI-720 (for southern chinch bug were allowed to probe, until any of them achieved recordings) or DI-710 (for western chinch bug re- stable (putative) xylem ingestion (H-I2 waveform; see cordings) analog-to-digital board (Dataq Instruments, later in the text). At that time, the insect would not be Akron, OH), as described by Almeida and Backus easily disturbed by movement. Three (rather than (2004), and recorded at 100 Hz by using a Dell Pen- four) insects were recorded because output from one tium notebook with WinDaq Pro software (Dataq). insect per day was recorded both pre- and postrecti- Experimental Design: Southern Chinch Bug. In the Þcation (Backus and Bennett 2009). Ri was then grad- three experiments performed, insects were attached ually switched through multiple varying levels (see to each of the four channels and were recorded for next paragraph) as H-I2 continued, to observe Ϸ20 h per day, under laboratory conditions and con- whether and to what degree the shape of this regular tinuous light. In the Þrst experiment, each channel continuing waveform changed with Ri level. All re- (i.e., insect) was set at a different Ri level with ap- cordings used DC applied signal (ϩ25 mV), so that propriate applied signal (for both AC and DC): Ri 106 native polarity of the output signal could be deter- ⍀ (ϩ30 mV), Ri 107 ⍀ (ϩ5 mV), Ri 108 ⍀ (ϩ5 mV), mined without offset adjustments (Backus and Ben- and Ri 109 ⍀ (0 mV). No negative applied signals were nett 2009). Thus, only gains were adjusted after each used. For the Þrst half of the insects recorded, the switch, for optimum display of waveforms, and then applied signal was initially DC, then was switched to recording was continued undisturbed for 15Ð30 s at AC at 6Ð7 h into the recordings, and then switched to each Ri level, between switches. The experiment was 0 mV for the last few hours. After the switch to AC, terminated as soon as the chinch bug was no longer offset and gain settings were adjusted for optimum performing H-I2. view of the waveforms (Backus and Bennett 2009). For 24 insects in the second experiment, EPG re- For the second half of the insects, the voltage was cordings began at Ri level of 107 ⍀. Switches were initially AC, then was switched to DC, and then attempted through Ri levels 106 (“R-only”), 107,108, switched again to 0 mV for the last few hours. In the 109, and 1010 ⍀ (the latter four Ri levels being sensitive second experiment, all four channels (i.e., insects) to varying intermediate combinations of R and emf) were recorded at Ri of 109 ⍀ with 2 mV applied signal. during each ingestion event. For an additional 27 in- Two channels were randomly assigned to AC applied sects, EPG recordings began with Ri 106 ⍀. Switches signal, and two to DC. In the third experiment, all made through Ri levels 106,109, and 1013 ⍀ (“emf- insects were recorded at Ri of 1013 ⍀ with a 0 mV only”) were again attempted. For 1013 ⍀, the applied applied signal. In total, 15 chinch bugs were success- signal was reduced to 0 mV before recording. Such a fully recorded in the Þrst experiment, Þve in the sec- large sample size of insects (51) was necessary be- ond, and three in the third experiment. Sample sizes cause it was seldom that all switches could be accom- of the last two experiments were small because Þeld- plished during a single ingestion event, or sometimes collected bugs became scarcer as the unexpectedly even a single insect, before the insect would be dis- hot and dry summer wore on. turbed and terminate ingestion. Nonetheless, with pa- During waveform analysis, it was found that certain tience, every Ri level was recorded at least six to eight aspects of waveform Þne- and medium-structure, es- times, with different insects. The exception was 1013 pecially relative amplitude, were highly variable ⍀, which only had two switching recordings, because among insects and Ri levels, probably because of dif- of the time-consuming amount of gain reduction ferences in wiring quality among insects (including needed after switching from 1010 ⍀. Amplitude of the use of silver print paint instead of the more conductive waveforms at 1013 ⍀ often exceeded display window silver glue). Consequently, to test whether waveform size (“peaked out”), because of a combination of ex- appearances could be more stable for each Ri level if treme ampliÞer sensitivity and the highly conductive, the wiring quality of each insect were improved large size of chinch bugs compared with the small size (through use of silver glue and thicker gold wire), a of most insects EPG-monitored previously. Gain ad- second set of experiments was planned. Because justments often could not be completed before the southern chinch bugs were no longer available, we insect became disturbed and terminated ingestion. decided to use western chinch bugs and also to com- Waveform Terminology. Waveforms were named pare waveform appearances between these two by following the hierarchical scheme established for closely related species. Homalodisca vitripennis (Germar) (formerly Homalo- Experimental Design: Western Chinch Bug. Wave- disca coagulata [Say]) sharpshooters (Almeida and forms were recorded in the same manner as for south- Backus 2004). Waveform phases describe broadly rec- ern chinch bug. Two experiments were performed ognized types of stylet penetration activities, such as under laboratory conditions with continuous light. In pathway versus ingestion, and can be seen at a coarse the Þrst, 14 insects were allowed to probe and perform level of resolution (WinDaq compression of 25Ð35). all behaviors (24 h/d) until stable sustained (putative) When the resolution is increased (WinDaq compres- phloem ingestion (J-I2 waveform; see later in the text) sion of 5Ð12), consistent patterns of waveform shape 528 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 4 and amplitude can be seen; these are termed wave- form families and labeled with capital letters. At a low level of compression (WinDaq compression of 2Ð5), the Þne structure of the waveform becomes visible and waveforms within a family are designated as types and given a number. Thus, waveform family-type com- bination is named by alphanumeric such as A1, A2, and so on. Each uninterrupted occurrence of a certain waveform type is termed a waveform event. In the case of regular repeating waveforms such as the I waveforms (see later in the text), one stereotypical

cycle of the waveform is termed an episode. Proposed biological meanings Waveform Analysis. Comparisons of Waveform Am- penetration of stylets to vascular bundle plitudes Across Ri Levels. Relative amplitudes for all of salivary ßange and trunk of salivary trunk salivation, tasting, testing of salivary seal into xylem tracheary element salivation into phloem sieve element element Secretion of thinner gelling saliva, deeper waveforms in southern chinch bug recordings were Secretion of thickest gelling saliva, formation Secretion of thickest gelling saliva, formation Penetration of phloem sieve element; Active ingestion from and possible salivation Active ingestion from xylemPassive/active tracheary ingestion element from and possible Passive/active ingestion from phloem sieve Salivation into vascular cell

estimated by identifying the highest absolute-ampli- ⍀ ⍀ ⍀ ⍀ ⍀ ⍀ ⍀ ⍀ ⍀ tude peak (measured from lowest valley to highest 7 7 10 9 9 9 9 9 9 Ð10 Ð10 Ð10 Ð10 Ð10 Ð10 Ð10 Ð10 Ð10 levels peak in that event) among all peaks in a waveform 6 6 7 7 7 7 7 7 7 Best seen at these Ri 10 10 event, and then estimating the amplitudes of other 10 10 peaks as a percentage relative to the highest peak R, emf 10 emf 10 emf 10 emf 10

voltage. Relative amplitudes were estimated from ϭ R, ϭ ϭ ϭ ϭ

each event of that waveform type in at least two ϭ probes made by each of Þve randomly chosen south- emf ϩ R, wave R, wave R, wave ern chinch bugs (total 16 events), and amplitude es- R, wave R ϭ ϭ ϭ ϭ timates were expressed as rounded percentage ranges. ϭ emf

Comparisons of Peak and Wave Amplitudes for ϭ Waveform H-I2 Across Ri Levels. A single representa- Electrical origin wave tive episode (termed a waveform excerpt) of waveform irregular family H-I2 (see later in the text) was visually selected Mixed; peaks and spikes as most representative of all repetitive episodes in each ⍀ H-I2 event recorded in Ri switches with western chinch and higher. 9 bugs. Excerpts were copied into an Excel Þle and ⍀ 9 Ri 10

grouped by Ri level and insect. A single excerpt then was extracellular Voltage level at Intracellular Mixed; rise and slope chosen to represent the average waveform appearance at Both intra- and that Ri level from all switching events, for each insect. Absolute amplitudes of the Þrst plateau (the “peak”), as ⍀ 9 well as the third and the eighth plateaus (in the “wave”; ϩ see also further description under Family I), were mea- Ri 10 waves 9Ð12 Hz sured, amplitude ratios for plateaus 1:3 and 1:8 were then underlying regular 7Ð9 Hz Rises irregular; calculated, and all ratios for all insects were statistically Irregular compared by Ri level. Measurement of Repetition Rates. Repetition rates ⍀ 7 of regular waveforms were measured in eight ran- ϩ domly chosen southern chinch bugs from the Þrst 7 9 ⍀ Ri 10 waves 9Ð12 Hz experiment (i.e., only Ri 10 and 10 ), and then underlying regular 7Ð8 Hz expressed as rounded frequency ranges. IrregularIrregular Irregular Irregular Intracellular Intracellular All R All R

Statistical Comparison Methods. All comparisons ⍀ a a were performed by using mixed model analysis of 9

variance (PROC GLIMMIX, SAS Institute 2001) fol- Ri 10 lowed by pairwise comparisons by using Fisher least ⍀ signiÞcant difference test. Data were not transformed 7 because homogeneity was strong according to Pear- Relative amplitude Repetition rate sonÕs ␹2 test. Results were considered signiÞcantly Ri 10 different at ␣ ϭ 0.05. G2 40Ð50% N/A H-I2 10Ð25% 70Ð90%J-I2 2Ð5% 6.5Ð9 Hz 5Ð10% 6.5Ð9 6.5Ð9 Hz Hz Extracellular 6.5Ð9 Mixed; Hz peak Intracellular Mixed; peak

Results I J-I1 2Ð5% 5Ð10% 6.5Ð9 Hz 6.5Ð9 Hz Intracellular Mixed; peak H 10Ð50% 30Ð70% Irregular General Overview of Chinch Bug Waveforms and Probing Behaviors. The coarse structures of both southern and western chinch bug EPG waveforms Not applicable; waveform family G is often not distinguishable from family H at Ri levels of 10 Phase Family Type clearly showed two main phases, termed putative Table 1. Summary of waveform characteristics for adult chinch bugs a J wave J 5Ð20% 20Ð40% Rises irregular; Pathway G G1 100% N/A pathway and ingestion phases (Table 1), as are always Ingestion I H-I1 10Ð25%Interruption 70Ð90% N 6.5Ð9 Hz 10Ð25% 70Ð90% 6.5Ð9 Hz Irregular Extracellular Mixed; peak Irregular Extracellular Extracellular 10 July 2013 BACKUS ET AL.: WAVEFORM LIBRARY FOR CHINCH BUGS 529 seen with salivary sheath feeders. Thus, chinch bug whereas negative waveforms could have peaks waveforms represented distinct phases of salivary pointed either negative-ward or positive-ward, de- sheath formation and stylet activities while searching pending on the waveform type. At Ri of 108 ⍀, voltage for vascular tissues (pathway phase), from which the levels could be either mono- or biphasic; 108 ⍀ voltage insects eventually ingest sap (ingestion phase). Two levels were almost always consistent within each in- shorter phases, the transitional J wave, and interrup- sect, suggesting that wiring quality was involved in the tions complete the chinch bug phases. Within wave- voltage level shift. form phase, waveform families G, H, I, and J were There also were no appreciable differences in ap- characterized, as well as Þner-structure voltage pat- pearance between southern and western chinch bug terns described below as individual waveform types. waveforms. Consequently, most waveform character- The waveform families occurred stereotypically in the izations described later are taken from the recordings sequence G 3H 3I 3H 3J 3I in most chinch bug of southern chinch bug, but apply to western chinch probes. bug recordings as well. Results of impedance (Ri) With the exception of waveform I (described later switching experiments with western chinch bugs are in the text), waveform appearances were similar be- incorporated into descriptions of waveform appear- tween Ri levels of 106 and 107 ⍀, as well as between ance, as relates to emf versus R components. All wave- levels 109 and 1010 ⍀. Waveforms recorded at Ri 108 ⍀ form characterizations are summarized in Table 1. were intermediate in appearance between those two, Waveform Families and Types. Family G. At the and Ri 1013 ⍀ waveforms were quite varied in appear- beginning of every probe, the Þrst waveform family, ance compared with all other Ri levels. Consequently, named G, was composed of a series of irregular-fre- representative Ri levels of 107 and 109 ⍀ were chosen quency very high-amplitude (tall) and rapid spikes or for the waveform Þgures, including both compressed narrow peaks (Fig. 1a). G family waveforms were and expanded views, to represent the evolution in divided into two types, G1 and G2 (Fig. 1b). G1 was waveform appearance across different Ri levels. composed of extremely rapid, irregular spikes of It was challenging to Þnd appropriate early probe mixed very high relative amplitude and voltage level, sequences that contained all waveform types for the always at the beginning of the probe (Fig. 1a and b). overview Þgures, because the sequences had to be G2 was composed of equally high-amplitude (but short enough to Þt in the space available for the Þg- slower) plateaus and valleys with multiple (high-fre- ures. Most H events in the G 3H 3I waveform quency) spikes on top of wide plateaus (Fig. 1a and b), sequence were typically several minutes in duration. termed “spikey plateaus.” G2 spikey plateaus always Most recordings would have required too much com- followed G1 spikes, and could lead to more G1 or pression and loss of detail to Þt in the Þgure space. The represent a transition to the H family (see later text). three probes with best lengths for display came from Individual G2 plateaus were often interspersed among two recordings that used different applied signal G1 events, or followed H-I (see later text). G wave- types, that is, AC for one probe at Ri 107 ⍀, and DC forms looked the same, regardless of whether the for two probes (from the same insect) at Ri 109 ⍀. Our applied signal was AC or DC. choice of waveforms for the Þgures is not intended to Both G1 and G2 were very tall at input impedance imply that these voltage types are optimum for re- (Ri) level of 106 or 107 ⍀ (Fig. 1a and b), whereas they cordings at the respective Ri levels. On the contrary, were shorter at Ri 108 ⍀ (not shown) and much we repeatedly found that when offset and gain settings shorter (often not distinguishable from H) at Ri 109 ⍀ were optimized for each Ri level (Backus and Bennett (Fig. 2a and b and c) or 1010 ⍀, and then disappeared 2009), there were no consistently detectable differ- entirely at Ri 1013 ⍀ (Table 1). Therefore, evidence ences in waveform appearance between AC, DC, and supported that the electrical origin of waveform fam- 0 V applied signal types, except for ingestion wave- ily G was strongly R-dominated with almost no emf. forms and Ri level 1013 ⍀ recordings, as described Voltage levels for G waveforms were strictly positive later. We did Þnd, however, that waveform appear- at Ri 106 and 107 ⍀ (Fig. 1a). In contrast, the base of ance was somewhat less variable among Ri levels when G (when present) was negative at Ri 109 ⍀ and higher, silver glue (rather than silver paint) and thicker wire but peaks extended positive-ward, so that the tops of were used. peaks were above 0 V (Fig. 2a and b). Two major changes in waveform appearances oc- Family H. Following G1 or G2, the amplitude of the curred as Ri level was changed. First, amplitude of waveforms distinctly decreased. This transition each waveform type increased or decreased, relative marked the start of waveform family H, composed of to each other, as Ri was progressively increased from overlying medium-structure rises and trenches that 106 to 1013 ⍀ (described for each waveform type). were highly variable and irregular-frequency, as well Second, voltage level of all waveforms shifted. At Ri of as underlying much smaller Þne-structure spikes and 106 and 107 ⍀, voltage levels were always monophasic, waves (Figs. 1a and b [s and w, respectively, in the that is, waveforms ßoated above the 0 V baseline, with inset boxes] and 2a, b, and d). H often represented the medium- and Þne-structure peaks pointed positive- major portion of the pathway duration in each probe. ward. At Ri of 109,1010, and 1013 ⍀, voltage levels were H waveforms looked the same, regardless of whether biphasic, that is, within a single probe, waveforms the applied signal was AC or DC, but its spikes and ßoated either above or below the 0 V baseline; positive waves were larger with 0 V applied signal than with waveforms always had peaks pointed positive-ward, either AC or DC applied signal. 530 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 4

Fig. 1. Representative southern chinch bug waveforms using 5 mV AC applied signal and 107 ⍀ input impedance (Ri). (a) Compressed overview of the early part of a typical probe with pathway followed directly by putative ingestion (H-I), then further pathway (H), followed by two J waves with subsequent, evolving J-I1. Waveforms are monophasic, with 0 V baseline at bottom of view. Windaq compression 300 (60 s/vertical div.), gain 4ϫ. Boxes b, c, d, e, and f are expanded in their respective Þgure parts. (b) Expanded view of box b in part a, showing G1, G2 “spikey” plateaus, and H pathway waveforms. Inset boxes show ampliÞed (Windaq gain 16ϫ) view, to reveal the small but distinct underlying wave (w) and 7Ð8 Hz-frequency spikelet burst (s). Inset compression 13 (2.6 s/vertical div.), Windaq gain 4ϫ. (cÐe) Expanded views of the same-lettered boxes in part a showing the evolution of putative ingestion waveform (H-I) appearance directly after pathway. Compression 2 (0.40 s/vertical div.) in this and all remaining Þgure parts. Windaq gain 8ϫ. (f and g) Expanded and ampliÞed (Windaq gain 32ϫ) views of the evolution of I appearance after a J wave; f is expanded from box f in part showing J-I1 10 min after the end of J; g is stable J-I2 from 4.3 h later in this probe (not shown in part a).

Relative amplitudes of Þne-structure H spikes were appeared that H itself was actually a complex under- not appreciably different at Ri levels of 106 through 109 lying waveform that existed simultaneously with G ⍀ (Figs. 1a and b and 2aÐc), yet they were quite high during early pathway, but was masked by the overly- and dominant in pathway at 1013 ⍀ (data not shown). ing and much larger G at low Ri levels. Thus, H had a At low Ri levels, H often appeared to alternate be- mixed electrical origin of both R and emf. Finally, at tween low-amplitude high-frequency spikelet bursts Ri 109 ⍀ or higher, the H waveform had a negative and much lower-amplitude hump-like waves (Fig. 1b, voltage level, thus was considered intracellular Þrst inset box; Table 1). At higher Ri levels, the waves (Fig. 2a). disappeared and H was composed entirely of spikelets Family I. Waveform family I began abruptly after of varying low amplitudes and frequencies (Fig. 2d, family H (or J waves; described later). I was a highly inset box). Thus, spikes and spikelets were emf-dom- regular-frequency (6.5Ð9 Hz; Table 1) low-to-high inated, whereas waves in humps were more R domi- amplitude waveform (depending upon Ri), occurring nated because they were visible only at lower Ri levels. at varying voltage levels (Figs. 1a [H-I1, H-I2] and 2a Also at very high Ri levels, H replaced G1 and some- [J-I1, J-I2]). The primary form of I was a series of times also G2 at the start of each probe; G was dimin- monophasic, narrow, rounded or peaked square waves ished in amplitude (relative to its height at lower Ri (termed rounded plateaus). Family I always began levels) (Fig. 2aÐc) or blended entirely with the then- with waveform type I1, which was highly variable in larger spikes of H (1010 and 1013 ⍀, not shown). It amplitude, as well as number and shapes of the pla- July 2013 BACKUS ET AL.: WAVEFORM LIBRARY FOR CHINCH BUGS 531

109 Ω Input Impedance (Ri)

episode wave

peak e f g

HGH-I1 evolving into H-I2 Z H J J-I1 evolving into J-I2

b f g e cd

j hi

a Fig. 6

b c d

h i

j k Fig. 2. Representative southern chinch bug waveforms using 2 mV DC applied signal and 109 ⍀ input impedance (Ri). (a) Compressed, side-by-side overview of the early parts of two typical probes. The Þrst excerpt shows pathway H followed directly by putative ingestion (H-I), then a short return to G then Z (baseline, nonprobing [also known as NP]). The second excerpt shows mostly H followed by a J wave, then J-I1 that evolves into J-I2. Both probes show biphasic voltage levels, with the gray dashed line indicating the 0 V baseline level (see text for further discussion). Compression 300 (60 s/vertical div.), Windaq gain 8ϫ. Boxes b, c, d, e, f, g, h, I, and j are expanded in their respective Þgure parts. (bÐd) Expanded views of boxes b, c, and d in part a showing a few G2-like plateaus, but for the most part, very few details in pathway H other than the underlying 7Ð9 Hz-frequency wave, difÞcult to further categorize. Compression 13 (2.6 s/vertical div.), Windaq gain 8ϫ. Unlabeled inset box in d shows expanded view to reveal the underlying 7Ð9 Hz-frequency wave. Compression 2 (0.40 s/vertical div.) in this and remaining Þgure parts. Windaq gain 8ϫ. Arrow in part d denotes baseline/Z. (eÐg) Expanded views of H-I waveforms in the same-lettered boxes in part a showing the evolution of putative I1 ingestion waveform appearance directly after H (Windaq gains 8ϫ). Boxes e and f show H-I1, g is H-I2. (hÐj) Expanded views of J-I waveforms in the same-lettered boxes in part a showing the evolution of putative ingestion waveform appearance directly after a J wave (Windaq gains 8ϫ for h, 64ϫ for i and j). Boxes h and i show J-I1; (j) shows J-I2. (k) J-I2 ingestion waveform, 2.2 h after part (j) (not shown on part a) (Windaq gain 64ϫ). 532 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 4

Fig. 3. Representative H-I waveforms of southern chinch bugs, using 107 ⍀ input impedance (Ri) and 0 V applied signal, showing complete evolution from H-I1 to H-I2 Þne structure. Large outer Þgure box shows overview of 190 s of H-I, starting immediately after H. Vertical dotted lines delimit interruptions (waveform N). Boxes a, b, c, and d are expanded in their respective Þgure parts. Compression 10 (2.0 s/vertical div.), Windaq gain 16ϫ. (a) Expanded view of box a, showing irregular peaks and waves of early H-I1. Compression 13 (2.6 s/vertical div.), Windaq gain 8ϫ, in this and remaining Þgure parts. (b and c) Expanded views of boxes b and c showing development of regular peaks and waves during mid-H-I1 and (in box b) interruptions (N). (d) H-I2, with stable Þne-structure that continued for another 174 s. teaus (e.g., Fig. 1f). I1 evolved in appearance over events differed dramatically from one another at high time, transitioning through highly variable intermedi- Ri levels (109 ⍀ or higher). H-I sequences always ate forms (e.g., Fig. 2e, f, h, and i). If the I waveform occurred at a positive (considered “extracellular” at Ri event lasted for more than a few seconds, repeated 109 ⍀) level, whereas J-I sequences always began at a episodes of plateaus developed. The start of each ep- negative (considered “intracellular” at Ri 109 ⍀) level, isode was a single plateau (termed the “peak,” Figs. 1e and either continued negative or gradually drifted and 2g) that was higher in amplitude than any of the toward positive, to about the 0 V level or barely pos- following plateaus within the episode (together itive (Fig. 2a). Thus, I waveforms following H were termed the “wave,” sometimes descending, as in Figs. distinctive in always being very large in absolute am- 1e, g, and 2g, and sometimes ascending, as in Fig. 2j and plitude, short in duration, and at a positive extracel- k). Episodic peak-plus-wave Þne-structure began to lular voltage level at Ri 109 ⍀; in contrast, I waveforms develop in mid-I1, eventually stabilizing at an appear- following J were always very short in amplitude, very ance that remained constant (i.e., stereotypical am- long in duration, and at a negative intracellular voltage plitude and Þne structure) within each Ri level for the level at Ri 109 ⍀. rest of the event (Figs. 1g and 2g, j, and k). An entire In addition, relative amplitudes of peaks in H-I2 sequence of this evolution is shown in Fig. 3. Once the (the Þrst plateau in each episode) decreased as the waveform appearance became stable peaks-and- Ri level was increased. This Þnding was demon- waves, it was termed I2. This “stable” I2 could con- strated by calculating the ratio of amplitudes of tinue, essentially unchanged, for many minutes to sev- plateaus 1 (the peak) versus 3 (in the wave), at each eral hours (1Ð3 h of I was often seen in long southern Ri level. The 1:3 amplitude ratio signiÞcantly de- chinch bug recordings). creased with increasing Ri (F ϭ 3.61; df ϭ 4,24; P ϭ I waveforms (including I1 transitional sequences 0.0192; Table 2; Fig. 4). As more plateaus occurred that eventually stabilized as I2) could occur at two in the wave in each episode, their amplitudes in- times during the stylet penetration sequence: 1) fol- creased relative to the height of the peak. Thus, the lowing pathway (usually after the Þrst H event, al- amplitude ratios of plateaus 1 versus 8 were not though sometimes only a very short-duration stretch signiÞcantly different across Ri levels (F ϭ 1.22; df ϭ of H after G2), hereafter termed H-I1 or H-I2 se- 4,24; P ϭ 0.3277; Table 2; Fig. 4). quences, or 2) following J (discussed below), hereaf- ter termed J-I1 or J-I2 sequences. At all Ri levels, H-I2 ϫ Table 2. Calculated peak ratios (means and SE) from H-I2 events were 5Ð10 (or more) taller in absolute am- episodes at different Ri (input impedance) levels plitude (Table 2) than J-I2 events (compare H-I vs J-I amplitudes in Figs. 1a and 2a). In addition, H-I wave- Ri Na 1:3 peak ratio 1:8 peak ratio form events (including both types I1 and I2) were 106 ⍀ 8 4.06 Ϯ 0.78a 5.23 Ϯ 1.26a always relatively short in duration (always Ͻ1h) 107 ⍀ 8 3.45 Ϯ 0.70ab 3.54 Ϯ 0.74a (Figs. 1a and 2a). In contrast, J-I events were very long 108 ⍀ 8 3.38 Ϯ 0.84ab 3.69 Ϯ 1.13a 9 in duration, usually exceeding 1 h, and often several 10 ⍀ 6 1.91 Ϯ 0.43c 2.29 Ϯ 0.73a 1010 ⍀ 8 2.59 Ϯ 0.66bc 2.96 Ϯ 0.88a hours long. Nonetheless, because episodes of both H-I 1013 ⍀ 2 0.99 Ϯ 0.81 1.06 Ϯ 0.87 and J-I were nearly identical in Þne-structure appear- ance and frequency (Table 1), it was logical that all I Different lowercase letters indicate signiÞcant differences among waveforms should be named the same, regardless of ratios in each column, not including Ri 1013 ⍀, for which the sample their amplitude and the waveform that preceded size was too small. a N ϭ number of insects from which one representative episode of them. stable H-I2 was measured. See Materials and Methods in text for Of interest, voltage levels of H-I and J-I waveform details. July 2013 BACKUS ET AL.: WAVEFORM LIBRARY FOR CHINCH BUGS 533

Fig. 4. Fine-structure waveform appearances of representative, stable H-I2 episodes of western and southern chinch bugs, at Ri levels of 106,107,109, and 1013 ⍀, using DC applied signal.

Finally, H-I2 episodes had a slightly different ap- tioned electrical evidence supports that H-I2 had pearance depending on whether an AC or a DC ap- mixed electrical origins. The wave portion of each I2 plied signal was used (Fig. 5). DC applied signal episode was primarily emf component, whereas the caused slight amplitude emphasis of the wave pla- peak was more R-dominated at lower Ri levels, until teaus, whereas AC applied signal caused slight em- Ri levels of 109 ⍀ or higher, when emf (especially the phasis of the peak at low Ri levels (Fig. 5). Also, the negative-going, biphasic peak; Fig. 2e and k) co-oc- peak at Ri 109 ⍀ and higher became intermittently curred with R. biphasic, with both positive- and negative-going por- Family J. An abrupt transition in waveform ap- tions, for most insects (Figs. 2k and 5). The aforemen- pearance often occurred at the end of waveform fam-

Fig. 5. Comparison of stable J-I2 waveforms from four different southern chinch bugs recorded simultaneously at four Ri levels (106,107,108, and 109 ⍀; note user annotations in traces). All four recordings were started with AC applied signal; recording was Þrst captured at Ϸ2.8 h after start of recording (left column) and 2.5 h after recording the J wave for the insect in channel 1. Insects on the other three channels had been ingesting since the start of recording. At 11.3 h after the start of recording, the applied signal was switched to DC. Recording was captured again (right column) at 15.5 h after the start of recording, when all four insects had achieved sustained J-I2 in subsequent probes at 2, 4.8, 1.1, and 3.0 h, after previously recorded J waves. Compression 2 (0.40 s/div.), AC Windaq gains were 128ϫ for all Ri levels, except 109 ⍀, which was 64ϫ; DC Windaq gains were 64ϫ for Ri 106 and 107 ⍀, and 128ϫ for Ri levels 108 and 109 ⍀. 534 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 4

Fig. 6. Southern chinch bug J waves. (a) Moderately expanded view of J wave and subsequent, evolving J-I1 waveform at Ri 107 ⍀. Boxes b, c, d, e, and f are expanded in their respective Þgure parts. Vertical bar with asterisk marks end of J wave with small but abrupt voltage drop. Compression 50 (10 s/vertical div.), Windaq gain 32ϫ. (bÐf). Expanded views of boxes b, c, d, e, and f in part a. Compression 2 (0.40 s/vertical div.), Windaq gains 64ϫ for (b), 128ϫ for (cÐf). (g) Moderately expanded view of J wave and subsequent, evolving J-I1 waveform at Ri 109 ⍀. Boxes h, i, j, k, and l are expanded in their respective Þgure parts. Compression 50 (10 s/vertical div.), Windaq gain 4ϫ. (hÐl) Expanded views of boxes h, i, j, k, and l in part g. Compression 2 (0.40 s/vertical div.), Windaq gains 16ϫ for (hÐj), 32ϫ for (k and l). ily H, often after a sequence of H 3I 3H events. At wave superimposed on the medium structure (Fig. 6b low Ri levels of 106 to 107 ⍀, sometimes also 108 ⍀,a and c). We deÞne a single episode of J wave as the medium-amplitude positive-going rise in voltage oc- duration from the voltage rise to the abrupt voltage curred. This rise was then followed by a complex drop (Fig. 6a and d); usually there were one or two stereotypically repeating waveform (termed a J episodes (Fig. 1a) before commencement of J-I1, not wave). The overlying medium-structure waveform more than two. consisted of a rapid positive slope to a medium-height, The overlying medium structure of the J wave was distinct, rounded peak (Fig. 6a and b), followed by a present with the same appearance (neither increas- relatively long negative-going slope (Fig. 6a and c), ing nor decreasing in amplitude) at all Ri levels from down to an abrupt end at a small voltage drop (Fig. 6a 106 to 1010 ⍀. At low Ri levels, where the general [*] and d). A much smaller underlying waveform also voltage level was monophasic, the overlying me- occurred, whose Þne structure was a low-amplitude, dium structure of the J wave was positive-going highly regular-frequency (9Ð12 Hz), peaked sine (Fig. 6a). However, at Ri 109 ⍀ and higher (and July 2013 BACKUS ET AL.: WAVEFORM LIBRARY FOR CHINCH BUGS 535 sometimes also at Ri 108 ⍀), when the general volt- “characterize”) the researcher-identiÞed wave- age level of the total recording became biphasic, the forms, and then to determine their biological mean- J wave inverted (Fig. 6g, h and j). The voltage drop ings (termed “correlate”). These experiments com- at the end of the J wave also inverted, becoming a prise the “triangle of correlations” (Walker 2000) to voltage rise. Yet the underlying, Þne-structure, deÞne the waveform via the insectÕs stylet activities short sine wave did not invert; it always remained (i.e., salivation, ingestion, stylet movements) and positive-going at all Ri levels. Interestingly, at Ri stylet tip location in the plant (e.g., mesophyll, 1013 ⍀, the overlying medium-structure curve of the phloem sieve elements, or xylem). The correlation J wave disappeared entirely, leaving only the un- process can be long and arduous, often requiring derlying Þne-structure sine wave that ended at a years of highly specialized methodology, such as still-present voltage drop. The aforementioned re- plant histology and light, confocal, or electron mi- sults indicate that the J wave medium-structure croscopy; excretory droplet studies; electromyog- curve is primarily R component. In contrast, the raphy; and other methods (Walker 2000). Nonethe- overlying Þne-structure sine wave is dominated by less, only after such work can the waveforms be emf, and the voltage change at the end is both R and deÞnitively deÞned and EPG knowledgably used as emf. Thus, the electrical origin of the J wave was a a tool for further studies. One of the largest chal- mixture of R and emf. lenges facing EPG science today is the pressing call Following the J wave, a highly variable but evolving to quickly adapt and use EPG as a tool for urgently J-I1 waveform always occurred. Its underlying Þne needed studies of insect feeding; this is especially structure began as a semi-regular (6.5Ð9 Hz), short, true for exotic and invasive hemipteran pest species. highly peaked square wave (Fig. 6e), gradually be- These days, however, there often is limited time or coming more organized into plateaus that slowly grew funding for the full, labor-intensive, and time-con- in amplitude and became more upwardly pointed with suming waveform correlation process. time. The evolution of this regular waveform was elab- The current study is an attempt to provide alter- orate and required many seconds to several minutes native methods to partially shorten the lengthy (Fig. 6e, f, j, k, and l). However, over time, the wave- waveform correlation and deÞnition process. Our form came to resemble the same I1 as typically fol- work builds on techniques developed over the past lowed H. The overlying medium structure was a 40 yr of EPG science (Tjallingii 1985, Backus 1994, gradual slope upward, eventually plateauing to a Walker 2000), but also uses novel capabilities of the more-or-less level constant-amplitude waveform (Fig. new ACÐDC monitor (Backus and Bennett 2009), 6a). Thus, the erratic but regular Þne structure of the especially its ßexible input resistor/impedance [Ri] immediate post-J I1 waveform evolved into appear- switches. Our objective was to learn enough about ances similar to H-I1 (compare Fig. 6k and l with Fig. the electrical properties of chinch bug waveforms to 2h and i), which within a few minutes evolved into hypothesize some important biological meanings typical I2 waveform (similar to Fig. 2j and k). Con- for waveforms, before completing the triangle of sequently, all I waveforms following the late-J wave correlations. We also sought to develop experiments voltage change are termed J-I1 evolving into J-I2 (Ta- that could become standards for future studies of ble 1). the electrical nature of waveforms. The current Family N. During H-I1 waveform, it was common study is not to suggest that electrical methods can, for short irregular interruptions of the waveform to or should, supplant traditional correlations. Com- occur. These interruptions, termed waveform family pleting the full triangle of correlation experiments N, took the form of an abrupt voltage rise, spikey for each new insect group is of great importance. leveling, then sudden voltage drop, resulting in a However, sometimes such tests can be exceptionally short-duration (a few seconds), wide, ßat-spikey pla- difÞcult to perform, as was the case with chinch teau (Fig. 3a and b). The voltage rise was typically to bugs. the same level as the nearby evolving H-I1 peaks (Fig. Although chinch bugs make salivary sheaths, some- 3). N interruptions always occurred during H-I1, be- times quite elaborate ones (Anderson et al. 2006, Ran- fore reaching the waveform appearance reaching sta- gasamy et al. 2009), they do so through the tightly bility in I2; they were never seen in J-I1. Interestingly, wrapped immature leaf sheaths in grass blades. Thus, the N plateau was always positive-going at low Ri each salivary sheath traverses multiple young leaves levels (106 and 107 ⍀), negative-going with larger separated by layers of air space. We made numerous more detailed spikes at the bottom at high Ri levels attempts to mark, excise leaf sheath tissues, and his- (109,1010, and 1013 ⍀), and either positive- or nega- tologically prepare salivary sheaths from correlated tive-going at 108 ⍀. EPG waveforms. However, the loosely wrapped leaf sheaths often fell apart, or salivary sheaths tore, during processing. In addition, we could not collect uncon- taminated excretory droplets, because chinch bugs Discussion prefer to lodge their bodies tightly between the leaf Whenever new insect species (such as the pres- sheaths and excrete within the tight space they ent two chinch bugs) are recorded for the Þrst time make. Consequently, we sought to thoroughly char- with EPG, a consistent series of experiments must be acterize EPG waveforms for chinch bugs, and then performed to Þrst identify and describe (termed to use electrical experiments to hypothesize the 536 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 4 putative biological meanings of these waveforms, phragm. Thus, for each I2 episode, perhaps waves sufÞcient to support tentative conclusions for future represent multiple cibarial pumps, each partially Þll- quantitative comparisons of resistant and suscepti- ing the cibarium (Dugravot et al. 2008), with peaks ble host plants. possibly representing complete opening and Þlling of The current study carefully described and char- the cibarium. Alternatively, the peak could represent acterized for the Þrst time the EPG waveforms of closure of either the cibarial diaphragm or the two related species of chinch bugs (Heteroptera: precibarial valve (blocking the precibarium, therefore Blissidae) and performed a rigorous study of the causing resistance to electrical ßow, that is, R), to electrical origins of their waveforms. EPG studies of swallow ßuid down the pharynx (Dugravot et al. numerous hemipteran species over the past 30 yr 2008). In a third alternative, I2 peaks may represent have allowed development of strong theoretical un- a brief spurt of salivation into a vascular cell. It is also derpinnings to describe the electrical origins possible that the higher amplitude of H-I wave- (termed R and emf components) of EPG waveforms forms, compared with J-I waveforms, means that (Tjallingii 1978, 1985; Walker 2000; Backus and Ben- greater suction is generated during cibarial pump- nett 2009). Of particular importance is understand- ing in H-I versus J-I. ing the R-emf responsiveness curve (Backus and Putative Phloem Contact, the J Wave. The J wave Bennett 2009), as well as the biological basis of R is a unique apparently species-speciÞc waveform and emf components. Information on the electrical that does not resemble any waveform previously origin of waveforms was combined with similarities published from any other EPG-recorded species. in appearance between chinch bug waveforms and Such a unique appearance is typical of an X wave, other species (aphids and several spe- an early concept in EPG science (McLean and Kin- cies such as sharpshooters). The result allowed us to sey 1967) that recently has been updated (Backus et hypothesize the following putative waveform deÞ- al. 2009). According to that recent article, an X wave nitions and biological meanings for chinch bug is “a multi-component waveform family represent- waveforms. These hypotheses can be tested in fu- ing multiple stylet activities that comprise tasting ture correlation studies. and testing of a possible preferred ingestion cell; Pathway Waveforms, G and H. The earliest wave- when preceding an ingestion waveform, (an X forms in stylet penetration undoubtedly represent for- wave) indicates [that subsequent] ingestion is from mation of the salivary sheath as stylets penetrate to a a preferred cell type such as phloem (aphids, leaf- preferred ingestion tissue (termed pathway activi- hoppers) or xylem (sharpshooters).” Accordingly, ties), as has been demonstrated with all other re- we propose that the chinch bug J wave is analogous corded salivary sheath feeders. Because saliva is highly to the X waves of aphids and . X waves electrically conductive, waveforms derived primarily are hypothesized to represent simultaneous pre- from salivation are heavily R-dominated. Chinch bug cibarial valve and cibarial diaphragm movements G2 strongly resembled the A1 salivation peaks of controlling minute ßuid uptake and egestion, com- sharpshooter leafhoppers (Backus et al. 2005), which bined with salivation (Backus et al. 2009). As an have been electrically and histologically correlated ecological concept, the X wave phase represents the with salivation of the sheath ßange and trunk, as well complex process of accepting a potential ingestion as stylet advancement. Similarly, chinch bug H resem- cell, analogous to acceptance of a host plant. Thus, bles the sharpshooter B1 waveform, which has been the X wave is a marker for ecologically important histologically correlated with stylet progression, nar- foraging behaviors, in any hemipteran. row sheath branch formation, and deep penetration to We propose that the overlying medium-structure the vascular bundle. Both A1 and B1 in sharpshooters curve of the J wave, highly R-dominated, represents are primarily R-component waveforms, similar to G the stylet tips pushing against the cell wall and/or and H in chinch bug recordings. membrane of a phloem sieve element, with the abrupt Putative Ingestion Waveform, I. All I waveforms voltage change at the end representing the tips break- have both R and emf origins; however, the most reg- ing through. Unlike for aphids, this membrane break- ular wave-like portions are emf-dominated, probably age is due to a mixture of R and emf, probably because because of streaming potentials caused by ßuid (either the large stylets of chinch bugs partially destroy saliva or food) pumping. Chinch bug I1 sometimes (rather than retain, as in aphids) the charge separation resembles the E1 waveform of aphids (Prado and of the cell membrane. The underlying, Þne-structure, Tjallingii 1994), especially at Ri of 109 ⍀. Therefore, I1 emf-dominated sine wave may represent salivation may represent salivation into a phloem cell, with or that begins as the stylets are pushed into the cell and without simultaneous passive ingestion. I1 evolves into continues within the phloem cell, analogous to I2 (regardless of whether preceded by H or J), whose E1 phloem salivation. peak-and-wave structure resembles the ingestion Putative Xylem- and Phloem-Ingestion Waveforms, waveform of corn leafhoppers, Dalbulus maidis Wol- H-I and J-I, Respectively. Given that chinch bugs are cott (Carpane et al. 2011). This peak-and-wave struc- considered predominantly phloem-ingesting species ture suggests active ingestion, that is, cibarial pumping (and assuming that our J-wave-equals-X-wave hy- of sap food. Dugravot et al. (2008) determined that the pothesis proves true, after future correlation studies), amplitude of sharpshooter ingestion plateaus was pro- it follows that J-I2 sequences may represent ingestion portional to the degree of uplift of the cibarial dia- from phloem sieve elements and H-I2 sequences may July 2013 BACKUS ET AL.: WAVEFORM LIBRARY FOR CHINCH BUGS 537 represent ingestion from xylem tracheary elements. environment. We also recommend use of gold wire This hypothesis is strongly supported by the electrical of diameter 25 ␮m instead of 12.7 ␮m, and use of Þndings in this article, especially that H-I is charac- silver glue instead of silver paint, for more consis- terized by: 1) positive, extracellular voltage level tent wiring quality. (characteristic for xylem ingestion waveforms in The present work characterized the EPG wave- aphids and sharpshooters), 2) high absolute amplitude forms for two species of chinch bug, western and suggesting higher cibarial diaphragm uplift required to southern, and found that they were similar. In ad- pump high-tension xylem ßuid, and 3) short durations dition, the current study provided electrical evi- of H-I sequences early in a probe (suggesting slaking dence for hypothesized biological meanings for all of “thirst” after dehydration during wiring). Similarly, of the chinch bug waveforms, including putative J-I is characterized by 1) negative intracellular voltage phloem and xylem sap ingestion. Evidence provi- level (characteristic for phloem-ingestion wave- sionally supports the conclusion that chinch bugs forms); 2) very low absolute amplitude (suggesting are salivary sheath feeders that are capable of in- minor cibarial uplift for barely-active, mostly passive gesting from both xylem tracheary and phloem sieve ingestion); and 3) very long durations of J-I sequences, elements, but preferentially from phloem cells. In later in the probe. the current study, chinch bugs performed a rather Putative Salivation Into and/or Tasting of Xylem stereotypical series of behaviors during stylet prob- Cells During N. Broad plateau-like interruptions oc- ing, starting with putative pathway activities (for- curred in early H-I1 waveforms, which strongly re- mation of a salivary sheath), progressing to short semble similar interruptions during sharpshooter events of putative xylem ingestion, followed by a recordings of xylem ingestion. Sharpshooter N return to sheath formation, and then an X wave-like waveforms are, like chinch bug N, positive-going at transition to phloem ingestion for variable long du- low Ri levels, but negative-going at high Ri levels rations. Further quantiÞcation of chinch bug feed- (Backus et al. 2005, 2009). Histological correlations ing, on both susceptible and resistant genotypes, have found that sharpshooter interruptions repre- will be possible now that the present waveform sent sheath salivation into and putative tasting of characterization has been accomplished. xylem cells (Backus et al. 2012). We propose a similar biological meaning for interruptions in chinch bugs. The absence of N waveforms following Acknowledgments J supports that J-I1 may be the sole form of salivation We thank Holly Shugart and Jose Gutierrez (U.S. Depart- in phloem cells. ment of AgricultureÐAgriculture Research Service) for their The aforementioned hypothesized biological mean- help during the EPG Workshop at California State Univer- ings of chinch bug waveforms will be useful for ten- sity, Fresno, in 2005, and for numerous subsequent attempts tative interpretations of quantitative studies compar- to collect excretory droplets and perform plant histological ing chinch bug feeding on different host plants. Such correlations for southern chinch bug waveforms. We thank Russell Nagata (UF/IFAS, Everglades Research and Educa- a study, comparing southern chinch bug feeding on tion Center) for supplying St. Augustinegrass. This research resistant and susceptible St. Augustinegrass, has been was supported by a Wedgworth fellowship (to M.R.) and the performed (Rangasamy 2008). Waveform quantiÞca- Florida Agricultural Experiment Station. tion has not been performed with data from the cur- rent study, because that has been done in the quan- titative study. References Cited Based on the results of the electrical experiments Almeida, R.P.P., and E. A. Backus. 2004. Stylet penetration described herein, it is recommended that future EPG behaviors of Graphocephala atropunctata (Signoret) studies of chinch bug feeding use 107 ⍀ as the standard (Hemiptera, Cicadellidae): EPG waveform characteriza- ampliÞer sensitivity (input impedance, or Ri). Ri of tion and quantiÞcation. Ann. Entomol. Soc. Am. 97: 838Ð 107 ⍀ provided the clearest visualization of all wave- 851. form phases, families, and types for chinch bugs. This Anderson, W. G., T. M. Heng–Moss, and F. P. Baxendale. was not the case for Ri of 106 ⍀, which often distorted 2006. Evaluation of cool- and warm-season grasses for resistance to multiple chinch bug (Hemiptera: Blissidae) or eliminated emf-dominated waveforms such as I, or species. J. Econ. Entomol. 99: 203Ð211. 9 ⍀ 10 and higher, which distorted or obscured R-dom- Backus, E. A. 1994. History, development, and applications inated pathway waveforms such as G. Either AC or DC of the AC electronic monitoring system for insect feed- applied signals can be used, as neither appeared to ing, pp. 1Ð51. In M. M. Ellsbury, E. A. Backus, and D. L. disturb the insects or reduce the information value Ullman (eds.), History, Development, and Application of of the waveforms recorded. (It should be noted, AC Electronic Insect Feeding Monitors. Entomological however, that very low voltage levels were delib- Society of America, Lanham, MD. erately chosen for this study. A few preliminary Backus, E. A., and W. H. Bennett. 1992. New AC electronic observations suggested that higher voltages [Ͼ100 insect feeding monitor for Þne-structure analysis of wave- forms. Ann. Entomol. Soc. Am. 85: 437Ð444. mV] of DC combined with low Ri levels [therefore Backus, E. A., and W. H. Bennett. 2009. The AC-DC cor- creating high current densities in the insect] caused relation monitor: new EPG design with ßexible input chinch bugs to cease probing.) As long as applied resistors to detect both R and emf components for any voltages are low, researchers can choose whichever piercing-sucking hemipteran. J. Insect Physiol. 55: 869Ð applied signal type leads to the least noise in their 884. 538 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 106, no. 4

Backus, E. A., M. J. Devaney, and W. H. Bennett. 2000. Eickhoff, T. E., F. P. Baxendale, and T. M. Heng–Moss. 2006. Comparison of signal processing circuits among seven AC Host preference of the chinch bug, Blissus occiduus. J. In- electronic monitoring systems for their effects on the emf sect Sci. 6: 6. (insectscience.org/6.07). and R components of aphid (Homoptera: ) Eickhoff, T. E., T. M. Heng–Moss, F. P. Baxendale, and J. E. waveforms, pp. 102Ð143. In G. P. Walker and E. A. Backus Foster. 2008. Levels of tolerance, antibiosis, and antix- (eds.), Principles and Applications of Electronic Moni- enosis among resistant buffalograsses and zoysiagrasses. J. toring and Other Techniques in the Study of Homopteran Econ. Entomol. 101: 533Ð540. Feeding Behavior. Entomological Society of America, Heng–Moss, T. M., F. P. Baxendale, T. P. Riordan, and J. E. Lanham, MD. Foster. 2002. Evaluation of buffalograss germplasm for Backus, E. A., J. Habibi, F. Yan, and M. Ellersieck. 2005. resistance to Blissus occiduus (Hemiptera: ). J. Stylet penetration by adult Homalodisca coagulata on Econ. Entomol. 95: 1054Ð1058. grape: electrical penetration graph waveform character- Heng–Moss, T. M., F. P. Baxendale, T. P. Riordan, L. Young, ization, tissue correlation, and possible implications for and K. Lee. 2003. Chinch bug-resistant buffalograss: an transmission of Xylella fastidiosa. Ann. Entomol. Soc. Am. investigation of tolerance, antixenosis, and antibiosis. J. 98: 787Ð813. Econ. Entomol. 96: 1942Ð1951. Backus, E. A., A. R. Cline, M. R. Ellerseick, and M. S. Serrano. Kerr, S. H. 1966. Biology of the lawn chinch bug, Blissus 2007. Lygus hesperus (Hemiptera: Miridae) feeding on insularis. Fla. Entomol. 49: 9Ð18. cotton: new methods and parameters for analysis of non- McLean, D. L., and M. G. Kinsey. 1964. A technique for sequential electrical penetration graph data. Ann. Ento- electronically recording aphid feeding and salivation. Na- mol. Soc. Am. 100: 296Ð310. ture 202: 1358Ð1359. Backus, E. A., W. J. Holmes, F. Schreiber, B. J. Reardon, and McLean, D. L., and M. G. Kinsey. 1967. Probing behavior G. P. Walker. 2009. Sharpshooter X wave: Correlation of of the pea aphid, Acyrthosiphon pisum. I. DeÞnitive an electrical penetration graph waveform with xylem correlation of electronically recorded waveforms with penetration supports a hypothesized mechanism for Xy- aphid probing activities. Ann. Enomol. Soc. Am. 60: lella fastidiosa inoculation. Ann. Entomol. Soc. Am. 102: 400Ð405. 847Ð867. Painter, R. H. 1928. Notes on the injury to plant cells by Backus, E. A., K. B. Andrews, H. J. Shugart, L. Carl Greve, chinch bug feeding. Ann. Entomol. Soc. Am. 21: 232Ð J. M. Labavitch, and H. Alhaddad. 2012. Salivary en- 242. zymes are injected into xylem by the glassy-winged Prado, E., and W. F. Tjallingii. 1994. Aphid activities during sharpshooter, a vector of Xylella fastidiosa. J. Insect sieve element punctures. Entomol. Exp. Appl. 72: 157Ð Physiol. 58: 949Ð959. 165. Baxendale, F. P., T. M. Heng–Moss, and T. P. Riordan. 1999. Rangasamy, M. 2008. Mechanisms of resistance to southern Blissus occiduus (Hemiptera: Lygaeidae): a chinch bug chinch bug, Barber (Hemiptera: Blissi- pest new to buffalograss turf. J. Econ. Entomol. 92: 1172Ð dae), in St. Augustinegrass. Ph.D. dissertation, University 1176. of Florida, Florida. Bonjour, E. L., W. S. Fargo, J. A. Webster, P. E. Richardson, Rangasamy, M., H. J. McAuslane, R. H. Cherry, and R. T. and G. H. Brusewitz. 1991. Probing behavior compari- Nagata. 2006. Categories of resistance in St. Augustine- sons of squash bugs (Heteroptera: ) on cucurbit grass lines to southern chinch bug (Hemiptera: Blissi- hosts. Environ. Entomol. 20: 143Ð149. dae). J. Econ. Entomol. 99: 1446Ð1451. Carpane, P., A. Wayadande, E. Backus, W. Dolezal, and J. Rangasamy, M., B. Rathinasabapathi, H. J. McAuslane, R. H. Fletcher. 2011. Characterization and correlation of new Cherry, and R. T. Nagata. 2009. Role of leaf sheath lig- electrical penetration graph waveforms for the corn leaf- niÞcation and anatomy in resistance against southern hopper (Hemiptera: Cicadellidae). Ann. Entomol. Soc. chinch bug (Hemiptera: Blissidae) in St. Augustinegrass. Am. 104: 515Ð525. J. Econ. Entomol. 102: 432Ð439. Cherry, R. 2001. Seasonal wing polymorphism in southern Reinert, J. A., and S. H. Kerr. 1973. Bionomics and con- chinch bugs (Hemiptera: Lygaeidae). Fla. Entomol. 84: trol of lawn chinch bugs. Bull. Entomol. Soc. Am. 19: 737Ð739. 91Ð92. Cline, A. R., and E. A. Backus. 2002. Correlations among AC Reinert, J. A., P. Busey, and F. G. Bilz. 1986. Old World St. electronic monitoring waveforms, body postures, and sty- Augustine grasses resistant to the southern chinch bug (Het- let penetration behaviors of Lygus hesperus (Hemiptera: eroptera, Lygaeidae). J. Econ. Entomol. 79: 1073Ð1075. Miridae). Environ. Entomol. 31: 538Ð549. Reinert, J. A., A. Chandra, and M. C. Engelke. 2011. Sus- Cook, C. A., and J. J. Neal. 1999. Feeding behavior of larvae ceptibility of genera and cultivars to southern chinch bug of Anasa tristis (Heteroptera: Coreidae) on pumpkin and Blissus insularis (Hemiptera: Blissidae). Fla. Entomol. 94: cucumber. Environ. Entomol. 28: 173Ð177. 158Ð163. Crocker, R. L., R. W. Toler, J. B. Beard, M. C. Engelke, and SAS Institute. 2001. PROC userÕs manual, version 6th ed. J. S. Kubicabreier. 1989. St. Augustinegrass antibiosis to SAS Institute, Cary, NC. southern chinch bug (Hemiptera, Lygaeidae) and to St. Slater, J. A. 1976. Monocots and chinch bugs: a study of host Augustinegrass decline strain of Panicum mosaic virus. J. plant relationships in the lygaeid subfamily blissinae Econ. Entomol. 82: 1729Ð1732. (Hemiptera: Lygaeidae). Biotropica 8: 143Ð165. Dugravot, S., E. A. Backus, B. J. Reardon, and T. A. Miller. Tjallingii, W. F. 1978. Electronic recording of penetra- 2008. Correlations of cibarial muscle activities of tion behaviour by aphids. Entomol. Exp. Appl. 24: 721Ð Homalodisca spp. sharpshooters (Hemiptera: Cicadelli- 730. dae) with EPG ingestion waveform and excretion. J. In- Tjallingii, W. F. 1985. Electrical nature of recorded signals sect Physiol. 54: 1467Ð1478. during stylet penetration by aphids. Entomol. Exp. Appl. Eickhoff, T. E., F. P. Baxendale, T. M. Heng–Moss, and E. E. 38: 177Ð186. Blankenship. 2004. Turfgrass, crop, and weed hosts of van Helden, M., and W. F. Tjallingii. 2000. Experimental Blissus occiduus (Hemiptera: Lygaeidae). J. Econ. Ento- design and analysis in EPG experiments with emphasis on mol. 97: 67Ð73. plant resistance research, pp. 144Ð171. In G. P. Walker July 2013 BACKUS ET AL.: WAVEFORM LIBRARY FOR CHINCH BUGS 539

and E. A. Backus (eds.), Principles and Applications of Walker, G. P. 2000. BeginnerÕs guide to electronic monitor- Electronic Monitoring and Other Techniques in the ing, pp. 14Ð40. In G. P. Walker and E. A. Backus (eds.), Study of Homopteran Feeding Behavior. Thomas Say Principles and Applications of Electronic Monitoring and Publications in Entomology. Entomological Society of Other Techniques in the Study of Homopteran Feeding America, Lanham, MD. Behavior. Entomological Society of America, Lanham, Vittum, P. J., M. G. Villani, and H. Tashiro. 1999. Turfgrass MD. insects of the United States and Canada. Cornell Univer- sity Press, Ithaca, NY. Received 12 February 2013; accepted 31 May 2013.